Are private schools really better? Not for everybody and not everywhere

Giuseppe Bertola 11 July 2013



Policy debates are frequent and emotional in many countries about voucher programmes meant to allow students who could not otherwise afford private schools to attend them. In May-June 2013, for example, controversies arose about such measures in the New Jersey, Wisconsin, and Pennsylvania State budgets, and about the results of a referendum on public funding of private pre-primary schools in Bologna, Italy. School vouchers are typically advocated by right-wing politicians who distrust governments’ ability to manage education (or anything else) and economists who favour competitive pressure and fear that heavily unionised public-school teachers may pay scant attention to the current and future needs of their students.

Voucher programmes might be a good thing

It is also possible, however, to view voucher programmes as an uncontroversial and sensible solution to a classic economic problem. If individual ability is complementary to school resources in producing education, then talented individuals are willing to pay for more and better educational resources than those supplied by a public school system that caters to the median voter’s ability (Stiglitz 1974). Then, publicly funded schools leave room for expensive private schools that, being attended by better students and using more and better educational resources, deliver better education. Relaxing the borrowing constraints that exclude brilliant but poor students from better schools, voucher schemes can improve equality of opportunities at the same time as they enhance the productivity of society’s educational resources.

Empirical evidence

This theoretical perspective fits empirical evidence from US, UK, and other Anglo-Saxon schooling systems, where private schools are attended by students who appear to be more talented as well as richer (Epple and Romano 1998; De Fraja 2002; Epple, Figlio, and Romano 2004). Educational resources can substitute rather than complement talent, however, so other configurations are possible. Students who enrol in private schools should be less talented than those of demanding government schools, and their school performance can be especially bad when private schools are no more expensive than government schools, but use their autonomy to cater to slow learners (Bertola, Checchi, and Oppedisano 2007; Brunello and Rocco 2008).

Figure 1. Country-specific differences between private and government school students’ average PISA test scores (in mathematics, reading, and science), plotted against country-specific differences between private and government schools’ percentage of total funding paid by parents

Bertola and Checchi (2013) find evidence in the 2009 PISA survey (OECD 2012) that private schools indeed do not everywhere deliver better schooling outcomes. Figure 1 shows that while in almost every country privately managed schools cover more of their costs with user fees than government schools, there is only a mildly positive association between the relative cost and apparent quality (as measured by the PISA test) of private education across all the 72 countries or regions in the data set. If the best students were selected into a private sector that supplies better education, PISA scores differentials should all be positive, and larger where the percentage of school funding from fees or charges paid by parents differ more across private and government schools. This is the case in all Anglo-Saxon countries (US, UK, Canada, Australia, and New Zealand), but in several countries (notably in Italy, and in Norway) private-school students do worse on the PISA test.

Detailed information collected by the PISA survey on individual school and students within each country makes it possible to study in more detail the qualitatively different roles played by private educational establishments in educational systems. School-level information is available on organisational aspects that indicate whether the school specialises in education of high-ability individuals, or in strengthening the educational outcomes of weak students. Differences in such respects between each country’s private and government schools can be interacted with indicators of each student’s ability to learn, and to choose private schools. The survey offers relevant information on the family’s wealth and cultural level, and the PISA test result itself (administered only a year or two after secondary-school enrolment) can be viewed as an indication of each student’s talent rather than of the school’s effectiveness.

Regressions that describe individual private-school enrolment choices detect a positive interaction between the family’s apparent wealth and the country-specific relative cost of private education, suggesting that ability to pay plays an important role in determining choice of privately managed schools. The family’s apparent cultural level, however, interacts negatively with indicators of country-specific academic performance expectation differentials across private and government schools. These and other results suggest that privately managed schooling is not necessarily chosen for students who, on the basis of their family’s cultural level and test performances, appear better able to cope with a demanding learning environment.

Further research

Further research should study the historical and political determinants of whether, in each country, privately managed schools are funded by the government (as is the case in Austria, Czech Republic, Finland, Germany, Hungary, Iceland, Latvia, Lithuania, Netherlands, Norway, Slovak Republic, Serbia, Slovenia, and Sweden), and whether the teaching organisation of government schools caters to low- or high-ability students, leaving different market niches to be filled by private schools. But existing evidence already indicate that voters and policymakers should not take it for granted that ‘private schools are better’. They are indeed better for students who choose to attend them, but may be seeking remedial education when government education is demanding, rather than opportunities to express their talent because government schools cater to low-ability students.

Policy implications

In countries with basic government-provided education, private schools are allowed to occupy a high-quality market niche. The policy menu should include improvement of public education standards as well as vouchers, which policymakers in countries with better public education should not adopt without considering their distributional and efficiency implications. If high-quality government schools attract the brightest segment of the student pool, then vouchers funding privately organised education benefit students who are not rich or dumb enough to purchase unsubsidised remedial education. While the resulting redistribution across differently wealthy and differently able individuals may be politically attractive in some cases, voucher schemes do not enhance overall equality of opportunities and efficiency in countries where governments supply high-quality education.


Bertola G Checchi D (2013), “Who Chooses Which Private Education? Theory and International Evidence”, CEPR Discussion Paper 9513.

Bertola G Checchi D Oppedisano V (2007), “Private School Quality in Italy", Giornale degli Economisti e Annali di Economia 66(3), 375-400.

Brunello G Rocco L (2008), “Educational Standards in Private and Public Schools", Economic Journal 118(533), 1866-1887.

De Fraja G (2002), “The design of optimal educational policies”, Review of Economic Studies 69: 437-466.

Epple D and Romano R (1998), “Competition between private and public schools vouchers and peer-group effects”, The American Economic Review 88(1), 33-62.

Epple D, D Figlio and R Romano (2004), “Competition between private and public schools: testing stratification and pricing predictions”, Journal of Public Economics 88, 1215–1245.

OECD (2012), "PISA 2009 Technical Report", Paris, Organization for Economic Cooperation and Development.

Stiglitz J (1974), “The demand for education in public and private school systems”, Journal of Public Economics 3, 349-385.



Topics:  Education

Tags:  private schools, school vouchers, private education

Professore ordinario, Università di Torino; CEPR Research Fellow


CEPR Policy Research